Precise definitions of grammar classes have always been a vexing subject. Bermudez and Schimpf demonstrate that the Not-Quite-LALR(1) grammars, long thought to be a strict superset of the SLR(1) grammars, are in fact neither a superset nor a subset and that the two classes are unrelated. In the process, they provide a clear description of the algorithms used to construct both SLR(1) and NQLALR(1) parsers.
The problem solved here is a minor one with little practical importance. However, many have fallen into the algorithmic error that gives rise to the NQLALR(1) grammars; perhaps we can hope that this paper will banish the error forever.